The psychometric performance of generic preference-based measures for patients with pressure ulcers

Interventions to prevent and treat pressure ulcers have been identified as a major cost for healthcare providers [1]. In the UK, it has been estimated that the cost of pressure ulceration to the UK health service could be up to 4 % of total expenditure [2]. The cost of treating pressure ulcers depends on their severity and category, and has been estimated as £1214 for the least severe (Category 1) and £14,108 for the most severe (Category 4) [3]. Despite these costs evidence for the cost effectiveness of treatment and prevention strategies is poor [4].

In order to undertake economic evaluations of treatments and prevention strategies for pressure ulcers one essential component is the ability to evaluate their impact on quality of life. This can be done using preference-based patient reported outcome measures (PROMs) which can be used to derive utility values that, when combined with mortality data, can then be used to generate Quality Adjusted life Years (QALYs) [5]. The UK National Institute for Health and Care Excellence (NICE) currently recommends that a generic preference-based measure (i.e. the EQ-5D) is used to generate the utility values for QALYs. However, there is evidence that generic measures may not be sensitive to the quality of life impacts of all conditions [6]. For some health issues, there are also preference-based condition-specific PROMs available: for example dementia [7], venous ulceration [8] and cancer [9]. However, there are a number of concerns with the use of condition specific preference-based PROMS. Firstly, the utilities derived cannot be easily compared (with each other, and with generic PROMs such as EQ-5D), and secondly the impact of comorbidities and/or side-effects may not incorporated into any assessment of quality of life [10]. Furthermore, condition specific preference based PROMs are not widely accepted by reimbursement agencies such as NICE.

The use of PROMs to evaluate the impact of pressure ulcers has been relatively limited [11]. One study compared SF-36 between patients in the community with and without pressure ulcers. They found that a large proportion of the patients were excluded as they were unable to consent, and there were difficulties completing the questionnaires [12]. Essex et al. [13] found that the SF-36 and EQ-5D were sensitive to the impact of pressure ulceration compared to those with no pressure ulcers within the acute setting, and a cohort study of hospital inpatients [14] also found that there were difficulties in recruiting participants and that there were poor completion rates using the SF-36. A condition-specific pressure ulcer PROM has recently been developed (PuQoL [15]). The PuQoL has a large number of questions (n = 82) covering ten scales: pain, exudate, odour, sleep, activities of daily living, emotional wellbeing, self-consciousness and appearance, and social participation. A seven dimension preference-based measure has been developed from the PuQoL [16].

In considering the best means of capturing the impact of pressure ulcers on quality of life there are two key considerations which may impact on the use of condition-specific measures such as the PuQoL.

Firstly, the likely domains of quality of life that will be affected for those with a pressure ulcer. Previous studies have highlighted that pain can be a major impact of having a pressure ulcer [17, 18]. Other studies have also described that having a pressure ulcer can have a psychological impact [19], affect usual activities [18], self-care [20] and social activities [20, 21]. Many of these domains are already incorporated into generic measures. One impact that has not been commonly included within either generic measures or the PuQoL is dignity. Having a pressure ulcer has been highlighted as resulting in a loss of self-perceived dignity [19, 22]. This may be particularly acute in those that have a pressure ulcer in the area of the sacrum [23] and for those patients reaching the end of their lives [24].

The second key issue results from pressure ulcers being a consequence of reduced mobility that maybe caused by a large number of diseases, medical interventions and co-morbidities [25]. Any disease or intervention that results in the patient being immobile or unable to change their own position for long periods puts the patient at risk of developing a pressure ulcer. Examples highlighted include: having surgery [26], dementia/cognitively impairment [27], stroke or neurological disease [28], spinal injury [29], diabetes [30], malnutrition [31] and even faecal incontinence [32, 33]. Therefore there may be the potential for any impact of pressure ulceration to be masked by the diversity and range of potential causes of pressure ulcers.

The variety of possible diseases and patient factors linked to pressure ulceration and the issue of impaired dignity being a major impact may mean that the most effective method of capturing the impact of pressure ulceration would be through the use of a generic rather than a condition-specific tool. In addition, studies that have explored the impact of pressure ulceration have highlighted that the key impacts are related to pain, usual activities, mobility, and social impacts, all of which are covered by generic measures. Other issues are directly related to the consequence of having an open wound–such as exudate and smell–but this may be captured within the wider scope of their influence on perceived dignity.

The aim of this study was to explore the psychometric performance of generic preference based PROMS (EQ-5D and SF-6D) in patients with pressure ulcers, or at risk of developing them, to examine the feasibility of using these measures in this patient group.

It is well established that pressure ulcers have a significant impact on quality of life [21]. In recent years, research has attempted to establish the best method to capture the impact on quality of life to better evaluate the clinical and cost effectiveness of interventions for the treatment and prevention of pressure ulcers. Pressure ulcers are a consequence of reduced mobility which may result from a wide range of both acute or chronic illness and treatment. The significance of this is that there are likely to be difficulties in separating the impact of the cause from the consequence (i.e. pressure ulcer) of the illness, and therefore generic measures which allow for the wider assessment of the HRQL impacts of conditions, may prove useful. Pressure ulcers are a consequence of being immobile or having reduced mobility. Therefore by definition anyone who has an intervention, disease or co-morbidity that affects mobility will be at risk of developing a pressure ulcer. The potential number of interventions and diseases may mean that having one condition-specific instrument able to capture the impact for this variety would be difficult. The development and use of PROMs specifically tailored to measure clinical consequences of a condition, such as pressure ulcers, may therefore be problematic: in the case of pressure ulceration has been described as “difficult, if not impossible” [20].

In the current study both the generic measures (EQ-5D and SF-6D) showed the ability to discriminate between those who were considered to be at risk of developing and those who had a wound, and also based on the severity of pressure ulcer. The presence of a category 1 to ≥3 as assessed for the acute population also led to a significant deterioration in score for both measures. These results suggest that generic PROMs can effectively capture the impact of pressure ulcers on quality of life, and therefore have some level of validity for use in groups of patients with pressure ulcers in both the acute and community sector. The results were in line with other work testing generic (EQ-5D and condition specific (PuQOL-UI) utilities in pressure ulcer patients [16]. Given the ease of completion of the EQ-5D and SF-12, the ability of generic measures to directly compare across conditions, and their likely sensitivity to the impacts of the underlying health condition leading to the pressure ulcer, generic measures can be used with some level of confidence in pressure ulcer patient groups.

The convergent validity results show that the utility scales are moderately correlated, but there are lower correlations between the descriptive systems, even across dimensions where a higher correlation would be expected (for example EQ-5D mobility and SF-6D physical functioning). This has been demonstrated in other health areas [43] and suggests that although the instruments are measuring some overlapping constructs, there are divergences that may be somewhat masked by the complementary nature of the utility scales.

The current study highlighted that collecting data from patients who have developed, or are at risk of developing, pressure ulcers can be achieved most effectively within the acute setting rather than the community. The large number of patients who were unsuitable or unable to participate in the study due to poor cognition and clinical condition also suggested that administration should be guided by interviewers rather than relying on self-completion. This has potential resource issues due to the increased cost of interviewer guided compared to postal administration. The difficulties of obtaining data on pressure ulcers within the community setting, due to the complexity and variety of localities, has also been highlighted previously [49]. It is also likely to be compounded by the nature of the population who develop pressure ulcers who tend to be elderly, in poor health and have potential cognitive impairment [50, 51]. Within the current study more than half the potential participants could not be approached due to problems with cognition or their clinical condition being too severe. In those who completed the survey 78 % had one or more comorbidities, 49 % had two or more, and 21 % had three or more comorbidities. The impact of an ill population and the selection bias imposed by only those able to answer questions would likely be even more of an issue in the community. However, data was successfully collected using both handheld devices and postal methods, and combining data collection modes may increase response rates provided that score equivalence can be demonstrated. There is evidence for this for the newly developed five level EQ-5D [43]. The recently developed EQ-5D-5 L may also show improvements in sensitivity for this group of patients.

This study has highlighted the complexity of those who develop pressure ulcers. It also emphasised that these wounds are a consequence of a range of conditions and interventions which result in reduced mobility. One consequence of this may be that a range of PROMs would need to be constructed to be sensitive to all of the HRQL impacts. However, the current study highlighted that existing generic measures seem to incorporate many of the important impacts of pressure ulceration. It may therefore be most appropriate to use generic rather than condition-specific PROMs when conducting economic evaluations of interventions aimed at preventing and treating pressure ulcers.